There’s no doubt SEO has evolved dramatically over the past year. With AI LLMs entering the chat, the way people are searching and the way brands are visible has fundamentally changed. As such, your search strategy should be reviewed too. Lucky we’ve done the groundwork for you: search hasn’t collapsed, but it’s evolved – and here’s what to focus on to sure up your strategy.
SEO’s found itself a new search partner
Conventional wisdom says AI overviews have triggered a wave of zero-click behaviour where users get the answer and never visit another site. And yes, the research holds up (*ahem* even more reason to refine your GEO strategy), but other data adds some important elements to the bigger picture.
Users are increasingly starting their research with AI LLMs, getting a synthesised answer, then heading to Google to verify it. It’s not AI instead of Google. It’s AI and then Google. A recent Semrush study backs up this surprising insight. When Semrush tracked some 10,000 keywords before and after an AI overview appeared, the zero-click rate decreased from 33.75% to 31.53%. Users aren’t stopping at the AI summary.
Meanwhile, Semrush’s research into Google’s AI Mode found that the average query was nearly twice the length of a traditional search (7.22 words vs. 4.0 words), which tells us that users are entering more conversational, complex questions, then using follow-up clicks to explore. In short, the search session has changed shape, not disappeared.
A key insight from Semrush’s AI Mode comparison study: domains that ranked well in Google’s organic top 10 results were far more likely to be cited in AI-generated answers, with a remarkably strong correlation at the domain level. Visibility in AI answers and visibility in traditional search rankings are both required as your potential customer might encounter you in either place, or both.
Backlinks: quality over quantity
The era of accumulating backlinks like a collector is over. An analysis of 35,000 data points by growth strategist Kevin Indig found that pages with high-quality, diverse links appeared in AI summaries 20% more frequently than those relying on quantity alone.
Companies are spending upwards of USD$1,000 per high-quality link acquisition, because one authoritative link from a relevant, trusted domain delivers compounding returns that dozens of low-authority links can’t match. For GEO specifically, link diversity matters too. Sites with a broader range of linking domains, including social platforms and directories, performed better in AI-generated responses. The algorithm is reading contextual signals that mimic how humans assess credibility.
Nofollow links matter
Here’s a widely misunderstood area of link strategy: nofollow links matter more than most people think. Since 2019, Google announced it would treat nofollow attributes as hints rather than hard directives, so although they don’t pass direct link ‘equity’, they may still pass value.
For GEO, the picture is even clearer. Indig’s analysis found that AI search engines actively factor in nofollow links when assessing domain authority and brand legitimacy. A brand mention or citation on a high-authority platform, even without a followed link, contributes to how AI models perceive your credibility. Press coverage, forum mentions, directory listings, and review platform entries all build the signal network that AI uses to form its understanding of your brand.
Investing more time into image backlinks
Most backlink audits focus exclusively on text links. But images carry their own link equity and AI search increasingly rewards multimodal content. When another site embeds your image and links back to your domain, that functions as a backlink.
When your images are properly optimised with descriptive, keyword-rich alt text and semantic file names, search engines gain a richer understanding of your page’s context. Google’s own documentation confirms it uses standard HTML image elements to find and index visuals, and that the metadata of the page where an image appears significantly influences how and where that image surfaces in search results. Similarly, the GEO search ecosystem increasingly synthesises answers from multiple content types including visuals.
The steps you do you need to take to help with image indexing are a bit finicky, but they’re important. Start with looking to create strong, sharable visuals such as infographics and apply descriptive alt text (not keyword stuffing), semantic file names, and add schema markup like ImageObject to give AI models richer context.
Final thoughts
Search in 2026 and beyond is not an either/or choice between SEO and AI. It’s a loop, as users move between AI answers and traditional results to form their understanding and make decisions. Brands that appear in both have a structural advantage.
The practical priorities: build fewer, better backlinks from topically relevant, authoritative sources; don’t dismiss nofollow links or image backlinks as secondary; invest in brand presence across trusted third-party platforms so AI systems encounter you in multiple credible contexts; and structure your content with schema, clear language and genuine depth – so that when AI synthesises an answer in your category, your voice is part of it.
